BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:WGAN and Optimal Transport - Mark Rowland\; Wenbo Gong
DTSTART:20171116T133000Z
DTEND:20171116T150000Z
UID:TALK94219@talks.cam.ac.uk
CONTACT:Alessandro Davide Ialongo
DESCRIPTION:*Abstract*\n\nOptimal transport metrics are methods of compari
 ng probability distributions\, and are increasingly used in machine learni
 ng. We'll give a general overview of these metrics\, and discuss some of t
 he statistical and computational issues with them when applying them in ma
 chine learning problems. The second part of the reading group will focus i
 n on the Kantorovich-Rubenstein duality of Wasserstein distance and the Wa
 sserstein GAN\, a generative model introduced earlier this year that claim
 s to fix some of the training difficulties associated with GANs\, using an
  optimal transport distance in its training objective.\n\n*Recommended Rea
 ding*\n\nThere is no required reading\, although the WGAN paper (http://pr
 oceedings.mlr.press/v70/arjovsky17a/arjovsky17a.pdf) contains useful backg
 round material.
LOCATION:Engineering Department\, CBL Seminar Room 4-38
END:VEVENT
END:VCALENDAR